The tetracycline challenge

The tetracycline challenge
18 May 2013 - Industrial chemistry

In what must be the single most important organic chemistry journal, each issue of Organic Process Research & Development features brave organic chemists doing brave organic chemistry on kilogram scale with chemicals that really matter. Take the guys from Tetraphase Pharmaceuticals who report on a grand-scale synthesis of a fluorinated tetracycline in the most recent issue (DOI) . It will come as no surprise the mission on the whole was a success but more interestingly is it to read about the setbacks and failures encountered on the way. That is what the journal is good at.
In this tetracycline story several challenges stood out. A simple Dieckmann condensation proved difficult because successful deprotonation in step one was found to depend on the quality of the particular batch of n-butyllithium. Remarkably in a trial with different BuLi batches the batch with the most solid floating around was most effective. This on account of LiCl present. Addition of Et3N.HCl neutralised the effect.
Next the scheme required a tert-butyldimethylsilyl group deprotection and efforts were made to have HF replaced by a more homonid-friendly alternative. Alas, many were tried (trifluoroacetic acid, p-toluenesulfonic acid, HF-pyridine, TBAF) but all failed due to non-performing conversion or purity. The HF had to stay.
Then more issues with commercial grade chemicals and consistency. How hard can it be to supply HCl/methanol with a consistent molality? Tetraphase's supplier apparently was not up to the job but it mattered greatly because not concentrated enough triggered premature precipitation of an intermediate in the next hydrogenolysis. As a workaround the team made their own HCl/MeOH from the concentrated acid.

Troubles not yet over but with the finish line in sight one of the final steps was an acid chloride acylation to a tetracycline intermediate with no less than 6 potential acylation anchors. Solubility of the reactant was found to depend on the reaction scale (how?) but by a stroke of luck at just the right pH the overacylated sites were reversed to the original state.

Final result: half a kilo of eravacycline ready for the next clinical trial.